Skip to content

Latest commit

 

History

History
16 lines (10 loc) · 2.51 KB

Goals.md

File metadata and controls

16 lines (10 loc) · 2.51 KB

Goals

Community

We, the members of the HPCng community, strive to create a diverse community of backgrounds, ideas, skill sets, and perspectives. Such diverse perspectives are needed to unite the currently disparate capabilities of HPC and enterprise to allow for running advanced computational and data analytics at scale.

Engaging with these respective communities will enable a greater understanding of requirements as well as existing technologies that could be used as part of this vision. This will facilitate cross-pollination to create technologies and solve problems across the ecosystem related to incompatible technologies and architectures between traditional Enterprise IT and HPC.

Collaboration

Referring back to the two party analogy in the abovementioned background, very few people intermingle between these parties, even when a single organization (like oil and gas, pharmaceuticals, aerospace, etc.) hosts both parties. Further, it could be said that the misalignment between HPC and enterprise is detrimental to both communities. On the other hand, there is value both can offer to each other by leveraging the best capabilities across the ecosystem. Thus, we strive to unite these parties and technologies where it makes sense, create new technologies and methodologies as needed, and build the appropriate bridges. We believe that the best way to build these capabilities is not by corporate direction, but rather led by a community of people that this directly affects, where everyone is a stakeholder.

Goals

Using the above metaphor, while the technical expertise between both parties is obviously quite specialized, it is important for these parties to benefit from each other's experience, capabilities, and technologies when applicable. There is an outcome here where some amount of infrastructure can be easily shared between enterprise use cases (AI/ML, inference, compute and data driven analytics, etc.) as well as HPC and computational use cases (simulation, modeling, rendering, prediction, analysis, research, and science).

“The convergence of AI, data analytics and traditional simulation will result in systems with broader capabilities and configurability as well as cross pollination.” —Dr. Al Gara, Intel Projects

We strive to develop a collection of open source, community driven projects that represent the next generation of high performance computing system infrastructure. As such infrastructure naturally spans beyond a single utility, the HPCng effort is split into several projects and focus areas.